Evaluating theories of drought-induced vegetation mortality using a multimodel-experiment framework

Area of Science:

• Plant physiology
• Ecology
• Climate change biology

Background:

• Understanding drought-induced tree mortality is crucial for predicting vegetation responses to climate change.
• Piñon pine-juniper woodlands are susceptible to drought, with observed mortality during precipitation reduction experiments.

Purpose of the Study:

• To simulate and understand the physiological mechanisms underlying drought-induced mortality in piñon pine-juniper woodlands.
• To compare model predictions with experimental data to refine vegetation models.

Main Methods:

• Utilized six models of varying scales (individual plant to global) with advanced representations of hydraulic and carbohydrate dynamics.
• Simulated physiology of Pinus edulis-Juniperus monosperma during a 5-year precipitation reduction experiment.
• Compared model outputs with empirical data to assess predictive power.

Main Results:

• All models predicted co-occurring hydraulic failure and carbon starvation in dying trees.
• The duration of severe hydraulic failure and carbon starvation, not absolute thresholds, was a better predictor of mortality.
• Limited carbon and water exchange at key plant interfaces (stomatal, phloem, below-ground) were associated with mortality.

Conclusions:

• Mechanistic, physiology-based models integrated with empirical data improve understanding and prediction of vegetation responses to drought.
• Model-data comparisons, especially when models fail, reveal underlying mortality mechanisms.
• Integrating mechanistic pathogen models into vegetation models could enhance simulation capabilities for ecosystem dynamics.